Sony Semiconductor Solutions transitioned from “IMX” to “LYTIA” recently. This move signals a change in the mobile imaging sector. It moves the focus from simple components to categorized experiences. The LYT-700 series sits at the center of this shift. This report analyzes two sensors in this family: the LYT-700 and the LYT-702.

Both sensors use the 1/1.56-inch optical format. This size is often called the “golden ratio” for balancing size and performance. However, they follow different engineering paths. The LYT-700 focuses on versatility and high frame rates for compact devices. The LYT-702 (related to the IMX921 architecture) prioritizes spectral fidelity and signal-to-noise ratio. Vivo’s Camera-Bionic Spectrum (VCS) technology heavily influences the latter.

Architectural Foundations

Understanding these sensors requires looking at their shared technology. The modern sensor is a multi-layered stack integrating optics, conversion, and processing.

The 1/1.56-inch Format

This format defines the “performance-class” main camera. It offers a surface area of 40-45mm². This is significantly larger than older 1/2-inch sensors. Yet, it avoids the massive lens volume required by 1-inch type sensors. This balance allows for bright apertures (f/1.8 to f/1.5) without severe optical penalties.

Stacked CMOS

Both models use Exmor RS stacked technology. This separates photodiodes from logic circuitry. The top layer handles light collection. The bottom layer manages processing. This vertical stacking enables complex features like DAG-HDR and All-Pixel AF without reducing the pixel capture area.

The Standard Bearer: LYT-700

The LYT-700 brings high-end features to a wider market. It succeeds the IMX890. It aims to provide flagship-grade autofocus and HDR to the sub-flagship tier.

All-Pixel Auto Focus

The LYT-700 uses All-Pixel AF. A single microlens covers a 2×2 grid of pixels. This differs from traditional PDAF where pixels are masked. The sensor compares light intensity from left and right pixels to calculate phase disparity. This allows 100% of pixels to aid in focusing. It improves speed in low light and texture-less scenes.

DAG-HDR and LBMF

This sensor uses Dual Analog Gain (DAG) HDR. It reads pixel charge with two gains simultaneously. High Conversion Gain lifts shadows. Low Conversion Gain preserves highlights. Combining these creates a raw frame with extended range. Additionally, Less Blanking Multi Frame (LBMF) reduces the time between frames. This allows for faster burst capture, resulting in sharper computational photography.

Research indicates a variant called LYT-700C exists. This version caps readout at 30fps. It reduces cost and thermal load for compact devices like the Motorola Edge 50 Neo.

The Specialist: LYT-702

The LYT-702 is a semi-custom component. It is effectively the successor to the IMX921. It appears primarily in the Vivo ecosystem.

Vivo Camera-Bionic Spectrum (VCS)

The defining feature of the LYT-702 is VCS technology. This alters the physical Color Filter Array (CFA). Standard silicon sensors see light differently than the human eye. They require strong filters and digital correction. VCS modifies the filters to align raw spectral intake with biological response.

This modification yields a reported 20% improvement in Signal-to-Noise Ratio (SNR). It also improves color restoration by 15%. The sensor requires less digital gain to achieve the same exposure. This results in cleaner images in low light.

Direct Comparison

The LYT-700 acts as a general-purpose performance sensor. The LYT-702 serves as a specialized high-fidelity sensor. The table below details the specifications.

Spectral Response Physics

The quantum efficiency (QE) of a sensor describes its ability to convert incoming photons into electrons. Standard Bayer filters often suffer from “crosstalk,” where red photons leak into green pixels.

The VCS Difference

The LYT-702’s VCS (Vivo Camera-Bionic Spectrum) filters are physically distinct dyes. They shift the peak sensitivity curves to closer mimic the human eye’s cones. Standard silicon is overly sensitive to infrared and pure green (530nm). VCS shifts the green response slightly towards blue and red, reducing the “digital look” caused by metamerism failure in artificial lighting.

In-Sensor Zoom Mechanics

Both sensors utilize a 50MP structure to offer “In-Sensor Zoom” (ISZ). This feature allows for a 2x lossless digital crop, providing an equivalent focal length of approximately 46mm-50mm without a dedicated telephoto lens. However, the execution differs based on the sensor pipeline.

The Remosaic Challenge

The native Quad Bayer layout groups pixels of the same color (e.g., four green, four red). To output a 12.5MP image at 1x, the sensor bins these groups. To achieve a 2x zoom, the sensor must “remosaic” the center 12.5MP crop, rearranging the color data back into a standard Bayer pattern. This process demands high computational throughput.

LYT-700 Approach: Speed

The LYT-700 prioritizes read-out speed during remosaic operations. It is designed to switch between 1x and 2x modes with minimal shutter lag, making it ideal for street photography where capturing the moment is paramount.

LYT-702 Approach: Color Fidelity

The LYT-702, often paired with advanced ISPs, applies VCS correction during the remosaic step. This ensures that even in the cropped state, the spectral response remains accurate, avoiding the “washed out” look that sometimes plagues digital crops in high-contrast scenarios.

Video Pipeline Architecture

Video performance separates these sensors more starkly than still photography. The demands of reading 50 million pixels 30 to 60 times per second generate immense heat and data throughput issues.

DOL-HDR vs. Staggered HDR

The LYT-700 typically utilizes Digital Overlap (DOL) HDR in video modes. This captures two frames (short and long exposure) in extremely rapid succession. While effective, it can introduce minor motion artifacts in fast-moving subjects. The LYT-700’s primary advantage is its support for high frame rates, capable of delivering 4K at 60fps with HDR enabled on supported platforms.

Rolling Shutter Mitigation

A critical metric for video is the sensor readout speed. Slow readout results in the “jello effect” where vertical lines slant during pans. The LYT-700 series improves upon the older IMX766 by reducing line readout times. The LYT-702, however, due to the complex VCS filtering, often relies on the ISP (like the V3 chip) to mechanically stabilize footage and correct rolling shutter artifacts computationally, rather than relying solely on raw sensor speed.

Optical Interface Dynamics

The interaction between the glass (lens) and the silicon (sensor) is critical. The 1/1.56″ format sits at a specific inflection point for mobile optics.

The Diffraction Limit

With a pixel pitch of 1.0µm (native), the LYT-700 series is diffraction-limited around f/4.0. Most mobile lenses are f/1.8, well within the safety zone. However, the Chief Ray Angle (CRA) becomes critical. If the lens is not designed specifically for the sensor’s microlens shift, users will see “shading” or purple vignetting in the corners.

Z-Height Constraints

The “Slim” Advantage of LYT-700 is engineered for the “Slim Mid-range” category. Its packaging allows for a slightly lower Z-height profile compared to the flagship 1-inch sensors. This makes it the preferred choice for foldable phones (like the Motorola Razr series) where every millimeter of thickness matters.

Data Throughput & Bandwidth

Processing 50MP images requires massive bandwidth. The sensor interfaces with the application processor via MIPI C-PHY or D-PHY lanes. Calculating this throughput helps understand why some phones overheat or disable features like 4K120fps.

RAW Data Semantics

For computational photography enthusiasts, the nature of the data output is crucial. The sensor does not output an image; it outputs a voltage map.

Standard Bayer vs. VCS De-matrixing

The LYT-700 outputs a standard raw stream that complies with general Android HAL (Hardware Abstraction Layer) definitions. This makes it compatible with third-party camera apps (like GCam ports) more easily. The ISP applies a standard 3×3 color matrix to convert the raw data to RGB.

The LYT-702 outputs “VCS Raw.” This data is spectrally shifted. If a standard de-matrixing algorithm is applied, colors will appear skewed (greens might look cyan, reds might look orange). The sensor mandates a custom “Color Correction Matrix” (CCM) located in the kernel of the OS. This proprietary lock effectively prevents the LYT-702 from being used optimally with generic camera software.

Competitive Landscape

The LYT-700 series does not exist in a vacuum. Samsung System LSI and OmniVision offer compelling alternatives in the 1/1.56″ class.

Ecosystem Integration

The choice of sensor dictates the design constraints of the final product.

Future Outlook

Sony’s LYTIA rebranding successfully tiers the market. The LYT-700 secures the volume premium segment. The LYT-702 locks in strategic partners. The 1/1.56-inch format will likely remain dominant for standard devices. Future iterations may adopt 2-Layer Transistor Pixel technology to boost dynamic range further.